What 2026 Already Knows
Seven structural trends the forecasts won’t touch
Grüezi und Gueti Nöis!
The gap between what’s announced and what’s delivered is widening across geopolitics, finance, and technology – and the beneficiaries are rarely the people told to just be patient.
1. Manufacturing isn’t coming home
The political narrative of industrial reshoring is collapsing. Kearney’s 2025 Reshoring Index fell 311 basis points while imports from Asian low-cost countries surged $90 billion. Domestic manufacturing output? Up 1%.
Intel’s €17 billion Magdeburg fab is cancelled. Micron’s New York complex won’t complete until 2045. TSMC’s Arizona Phase 2 slipped to 2028. Europe’s chip share is falling from 8% to 6%, not rising to the target 20%.
Stanford research shows where production actually went: US imports from China down five percentage points, but China’s exports into Vietnam up five points, into Mexico up two.
The supply chain didn’t shorten – it added stops. Vietnamese and Mexican facilities remain tightly integrated with Chinese upstream suppliers.
Who benefits: Politicians banking announcement headlines; friend-shoring nations absorbing capacity without domestic policy costs.
Who loses: Taxpayers funding projects that stretch to the 2040s; anyone expecting supply chain security rather than its performance.
2. Insurance execs are climate warriors
The insurance industry is forcing climate adaptation at a pace political systems cannot match.
Federal Reserve Chair Jerome Powell back in February 2025:
“If you fast forward 10 or 15 years, there will be regions of the country where you can’t get a mortgage, there won’t be ATMs, banks won’t have branches.”
Munich Re warns of properties becoming “uninsurable, unsaleable, and ultimately unusable.” Lloyd’s directors say by the time climate shows definitively in claims data, “it will be too late.”
Across the United States:
Florida: 15+ insurers declared insolvent since 2020, premiums averaging $15,000 annually – five times the national average.
California: State Farm and Allstate stopped writing new policies; the FAIR Plan needed a $1 billion bailout after January’s wildfires.
Two million policies cancelled in five years, four times the normal rate.
$13 trillion in US mortgage debt depends on insurance availability. When insurers withdraw, mortgage markets contract, property values collapse, tax bases erode – a cascade no elected official can acknowledge.
Government subsidies step in where markets fear to tread. The US National Flood Insurance Program charges 35% of actuarial value, socialising risks that insurers are refusing to carry, quietly relying on taxpayers to pick up the pieces.
Who benefits: Those exiting climate-exposed assets early; real estate speculators in “climate haven” regions.
Who loses: Homeowners unable to move discovering stranded assets; communities facing 50% higher wildfire vulnerability.
3. Smart money’s exit
Lending migrated from regulated banks to private credit – and pension funds are holding the risk. The market grew from $46 billion in 2000 to $1.7 trillion today. Global shadow banking assets exceed $74 trillion.
Remember the IMF’s April 2024 Global Financial Stability Report?
“If private credit remains opaque and continues to grow exponentially under limited prudential oversight, these vulnerabilities could become systemic.”
More than a third of borrowers now have interest costs exceeding earnings. Average leverage in the US middle market: 7x EBITDA.
Pension funds hold 30% of private credit assets. The 200 largest defined-benefit funds increased exposure 57% in a single year. Sovereign wealth funds are the “sophisticated” buyers. Retail investors are the newest entrants.
Banks now lend to private credit funds through warehouse lines – creating contagion channels whilst claiming their balance sheets are clean. In the polite words of the Federal Reserve, the connections are “unclear.”
The losses won’t be: pension beneficiaries and retail investors who followed “smart money” into the dark.
Who benefits: Alternative asset managers capturing $28 billion+ in fees; partners extracting returns before the cycle turns.
Who loses: Retirees with pension exposure to unpriced risk; anyone expecting monetary policy to work through channels that no longer exist.
4. Not customers, partners
AI governance is framed as US-China competition. But Gulf sovereign wealth is now inside American AI infrastructure, not purchasing access to it.
The UAE’s MGX is one of four equity funders for OpenAI’s $500 billion Stargate Project and anchors a $100 billion infrastructure fund with BlackRock and Microsoft. Saudi Arabia’s PIF has signed $14 billion with NVIDIA, $10 billion with AMD, $10 billion with Google Cloud. The combined committed capital is approaching a quarter-trillion dollars.
Data centres need immense power. Gulf states offer cheap electricity, abundant land, and capital – concentrated resources that neither the US nor China can match.
Export controls created a market for “trusted intermediary” hosting: Gulf states can provide Global South countries US-stack infrastructure without direct chip licences. They’re positioning as backends for AI across emerging markets.
Saudi Arabia hosted Global AI Summits in 2020 and 2024. The GCC adopted OECD AI Recommendations whilst hedging – Saudi invested $400 million in China’s Zhipu AI alongside its US partnerships.
Who benefits: Gulf funds gaining leverage in AI governance; tech giants accessing cheap power.
Who loses: Anyone assuming AI rules will emerge from anything other than capital deployment.
5. Who trains the next partner?
Professional services firms are collapsing entry-level hiring while expecting expertise to materialise anyway.
PwC cut graduate intake 32%. UK accounting listings fell 44% year-over-year. EY delayed start dates three years running – 2025 hires begin March 2026. Law firm associate ratios dropped from 45% to 40% over fifteen years. Investment banks are testing AI that could eliminate two-thirds of junior analyst roles.
OpenAI’s Mercury project has 100+ ex-bankers training models on financial modelling, due diligence, and pitch decks – the work that once trained future partners.
Entry-level tasks aren’t make-work. They’re an apprenticeship. Harvard economist David Deming: “AI is best suited to tasks often assigned to junior professionals.”
And those tasks are how knowledge passes between generations. Wall Street Oasis: “If they don’t need the drudgery anymore, banks might be a little more reluctant to provide the apprenticeship.”
Who becomes senior partner when today’s juniors never learned the craft?
Who benefits: Partners near retirement, extracting cost savings before consequences kick in.
Who loses: Graduates facing closed pathways; clients in a decade when expertise hasn’t been replenished.
6. BRICS without mortar
BRICS expansion is positioned as an alternative order. The institutions tell a rather different story.
Jim O’Neill, who coined the acronym, in 2024:
“Each year brings further confirmation that the grouping serves no real purpose beyond generating symbolic gestures and lofty rhetoric.”
The New Development Bank has approved $40 billion since 2015; the World Bank lent $209 billion to Africa alone over a similar period. The NDB hasn’t lent to Russia since 2022 because of sanctions risk.
The BRICS Contingent Reserve requires IMF programme compliance for meaningful borrowing, binding members to the very institution they claim to oppose. 78% of NDB financing remains dollar-denominated.
Meanwhile China-India tensions persist since Galwan Valley. Saudi Arabia hasn’t formally accepted membership. Brazil pushed back against China’s expansion agenda.
The actual beneficiaries are intermediary states and mobile capital. UAE attracted 9,800 millionaires in 2025 – global leader. Dubai holds $550 billion+ in financial assets with zero income, capital gains, or inheritance tax. Turkey explores BRICS partnership while conducting 65% of oil imports from Russia.
Regulatory arbitrage is enabled by fragmented authority – not new rules exactly, but the absence of enforceable ones.
Who benefits: Mobile capital optimising across jurisdictions; intermediary centres extracting positioning rents.
Who loses: Fixed populations expecting development finance; anyone assuming multipolarity means accountability.
7. Waiting for gigawatts
AI’s binding constraint isn’t capital or talent – it’s grid connections.
US data centre power demand will triple by 2030. Those data centres will consume more electricity than all US energy-intensive manufacturing combined. The IEA projects global demand equivalent to Japan’s entire current energy consumption.
Meantime, data hub northern Virginia faces seven-year grid connection delays. Ireland imposed a moratorium after data centres reached 22% of its national consumption. All six major US grid operators missed their July 2025 FERC deadline. A US Department of Energy report warns blackout risks could increase 100-fold.
The nuclear response is real but late. Microsoft signed a 20-year PPA to restart Three Mile Island. Google agreed to buy power from seven SMRs by 2035. But only three SMRs exist globally, and none in the US. NuScale’s Idaho project was cancelled after costs tripled.
Meanwhile, inference costs collapsed 280-fold in two years. DeepSeek runs 20–50x cheaper than OpenAI equivalents.
Rapidly-depreciating models may not justify infrastructure that takes a decade to build.
Who benefits: First-movers with behind-the-meter power; natural gas as bridge fuel.
Who loses: Latecomers in grid queues; ratepayers funding upgrades; anyone expecting AI timelines decoupled from physical constraints.
The uncomfortable forecast for 2026 is not that these trends will reverse – it’s that those bearing the costs will remain underrepresented in the calculations that shape elite consensus.
Thanks for reading!
Wishing you health, wealth, and happiness in 2026.
Best,
Adrian